永磁球形电机Stribeck摩擦模型参数辨识

Stribeck friction model parameter identification for a permanent-magnet spherical motor

随速度变化的摩擦转矩使永磁式球形电机在跟踪控制过程中出现爬坡或抖动现象,影响跟踪控制精度且难以检测。结合Stribeck摩擦转矩模型提出一种基于粒子群算法(PSO)的参数辨识方法,建立起精简、准确的电机摩擦转矩模型。在电机动力学微分方程基础上加入摩擦转矩项,通过恒速摩擦转矩检测实验得到转速-摩擦转矩间的稳态对应关系;采用PSO算法对转速-摩擦转矩序列进行参数辨识,得到Stribeck摩擦转矩模型;再通过实验对辨识得到的摩擦转矩模型进行验证。研究结果表明,参数辨识得到的摩擦转矩模型与实验数据相比,误差控制在2.5%以内,可以证明所提出参数辨识方法的有效性。

The problem of the friction torque that changes with the speed causes the permanent magnet spherical motor to climb or shake during tracking control process, which affects the tracking control accuracy and is difficult to detect. In order to establish a simplified and accurate motor friction torque model, a parameter identification method of Stribeck friction torque model based on particle swarm optimization(PSO) is proposed. On the basis of the differential equation of motor dynamics, the friction torque term was added, and the steady-state correspondence between speed and friction torque was obtained through constant speed friction torque detection experiments. PSO was used to identify the parameters of the rotation speed-friction torque sequence to obtain the Stribeck friction torque model. The experiments verified the recognition effect of friction torque model. The research results show that the error of the friction model obtained by parameter identification is controlled within 2.5% compared with the experimental data, which can prove the effectiveness of the proposed parameter identification method.